NVIDIA

Mirantis and NVIDIA are partnering to make it faster and easier for developers to build and run GPU-accelerated containers using Mirantis Kubernetes Engine. The new integration is based on the NVIDIA GPU Operator, which automates the lifecycle management of the various software components required to expose GPUs on Kubernetes. It enables advanced functionality and increased GPU performance, utilization, and telemetry. Mirantis has worked with NVIDIA to validate the NVIDIA GPU Operator on Mirantis Kubernetes Engine 3.5.7+ and 3.6.2+, optimizing the GPU Operator to quickly provision the servers with GPUs and run containerized AI/ML workloads.

Mirantis OpenStack for Kubernetes also supports NVIDIA time-sliced vGPU and Multi-Instance GPU (MIG) modes for dynamic allocation of GPU resources or more isolated environments for predictable performance and better resource isolation.